Post on 16-Oct-2021
Int j simul model 19 (2020) 3, 375-386
ISSN 1726-4529 Professional paper
https://doi.org/10.2507/IJSIMM19-3-517 375
DISCRETE EVENT-BASED RAILWAY SIMULATION
MODEL FOR ECO-EFFICIENCY EVALUATION
Amorim, G. A.; Lopes, L. A. S. & Silva Junior, O. S.
Instituto Militar de Engenharia, Seção de Fortificação e Construção, Rio de Janeiro, RJ, Brasil
E-Mail: gracianoamorim2014@gmail.com, laslopes@uol.com.br, orivalde@ime.eb.br
Abstract
Based on the discrete event-based simulation method, this paper aims to model the Paranaguá “KM5”
Railyard, looking at the inbound and outbound movement of trains and wagons from an ecologic
perspective. The originality of this paper comes from the inclusion of the train parameters and eco-
efficiency concerning train operation. Moreover, it helps the operator to define a range of optional
strategies for wagon classification, which would be more environmentally friendly, also providing a
basis for further research on this theme. This research method is quantitative, based on visits to the
railyard, interviews with rail operators, the original railyard’s geometrical design, real operation charge
for 2016, as well as theoretical studies. The main findings are that the analysis of the simulated cases
can be applied by the rail manager to the railyard’s operation manual. The implication for theory and
practice is that the findings can also contribute, with different parameters of decision, not only the
reduction of time taken but also with consideration of the environmental impact. (Received in April 2020, accepted in July 2020. This paper was with the authors 1 month for 2 revisions.)
Key Words: Railyard, Discrete Event-Based, Simulation, Eco-Efficiency, Anylogic, Paranaguá
1. INTRODUCTION
Simulation is the technique of creating, in a virtual environment, a representation of a real
system using mathematical models to study the behaviour under normal operating conditions;
or to test hypotheses with the security of performing these tests without risks and with low cost
involved. The concept of simulation has been spreading with increasing speed in recent years
with the emergence of new, more robust, and versatile software [1].
The simulation methods can be divided into three categories: Monte Carlo, where events
are randomly generated and in large volume; Continuous Simulation, which analyses situations
where continuous changes occur over time; or Discrete Events, an adopted method in this paper,
where changes occur at certain moments [2].
Railyards involve complex asset management, with high operating costs and environmental
impact. In this sense, some authors have developed models to evaluate the environmental
efficiency of projects in general, such as the World Bank studies [3], as well as their adaptation
to freight locomotive analysis [4]. These studies allow the evaluation and comparison of various
rail transport solutions by eco-efficiency indicators concerning global references.
Fan et al. [5] propose a simulation method to evaluate systems of abrasive method grinding,
aiming for more efficient processes to remove defects from rails, such as corrugation, crack,
spalling and squat, prolonging the rail´s service life. The authors adopted the system SIMPACK
software to verify the performance in terms of dynamics indices (lateral and vertical vibration
accelerations, axle transverse force and derailment coefficient). The dynamic simulation model
carried out by SIMPACK, allows investigation into the performance of the designed device,
showing the variation of dynamic indices in terms of speed and geometry characteristics.
Wang and Chen [6] propose a simulation method of production logistics, analysis of results
and identification of bottleneck using the FLEXSIM system. By this model, they proposed
optimization of methods, reducing of waste and improving efficiency. The results proved that
the efficiency of the logistics could be improved using simulation modelling.
Amorim, Lopes, Silva Junior: Discrete Event-Based Railway Simulation Model for Eco-…
376
Chen [7] proposes a model to optimise the production planning and scheduling, considering
the logistics constraints. The author adopted the QPSO algorithm, which allowed resource
maximization, greater efficiency and bringing benefit to the enterprise.
Suryani et al. [8] propose an urban transportation planning using the system dynamics
simulation model. The objective was to optimise traffic management regarding congestion
mitigation, improving urban mobility. By applying intelligent mobility modelling, the authors
tested and evaluated strategies to mitigate traffic jams in the city of Surabaya, East Java,
Indonesia.
Fragapane et al. [9] propose a study to simulate the logistics in hospitals. Through modelling
of different scenarios, they can provide planners with support assistance, using the automated
guided vehicle (AVG), which analyses the internal demand for goods, improving its delivery
performance. The vehicle is operated by a radio-frequency communication system which
allows the opening of doors and calling of elevators.
The objective of this paper is to propose a discrete event simulation model for a railyard
and to compare the eco-efficiency of two classification strategies of wagons. The model
considers the operations of train arrivals, wagon organization and classification, crossing
control, wagons releasing to the port terminal, return to the railyard and sending back to their
origin.
In addition to this introductory section, Section 2 presents the research methodology.
Section 3 summarizes the theoretical concepts. Section 4 describes and analyses the Paranaguá
“KM5” Railyard simulations. Section 5 presents the conclusions and directions for future
works.
2. RESEARCH METHODOLOGY
This research is quantitative and descriptive, with a case study carried out in a “KM5” railyard
placed in the Paranaguá City, Paraná, Brazil. The investigations were concentrated in the yard
operations during the period June to November (2016), based on train scheduling of the
company responsible for the yard in that period. For this, the design of the railyard, the
worksheet with the schedule of trains arriving and leaving the yard, the description of the
wagons, products and customers were obtained. Site survey and interviews were conducted
with operation team technicians to understand how the yard is operated, how wagons are
classified, in which lines they are parked, and the main problems they have faced. Using the
results of the interviews and the site survey, it was possible to develop the simulation model of
the railyard.
3. THEORETICAL FOUNDATION
Several scholars’ research projects have been developed to study simulation models for specific
rail transport systems, both passenger and cargo. Among these studies, it is possible to highlight
the one carried out by Confessore et al. [10], the study regarding the railroad linking the cities
of Verona and Brennero, Italy. The authors proposed the simulation and optimization to
estimate the capacity of the pathway. The obtained results allowed a significant increase in the
vehicle flow by reducing the train’s interval, the environmental impact, energy consumption
and improvement of the use of railway infrastructure, with the composition of freight and
passenger trains sharing the same rails.
Crane et al. [11] developed a Monte Carlo simulation model for a freight railyard.
According to the authors, these studies are important since freight wagons generally spend 2/3
of their time parked in the yards or awaiting the classification of transported products. This
Amorim, Lopes, Silva Junior: Discrete Event-Based Railway Simulation Model for Eco-…
377
brings inefficiency to the system, increased fuel, and environmental impact. Through the
studies, the authors obtained positive results in reducing the total waiting time of the system.
The importance of simulation analysis in freight railways is due to the use of these systems
for operational optimization. Railways in the USA accounted for 35.6 % of cargo traffic [12].
Despite this, old-fashioned processes for increasing operational efficiency and processes
rationalization are still adopted. Assad [12] describes several studies that have been developed
in simulation and operational optimization, concluding their great advantage and how the
development of new technologies could increase the scope of these analyses and the efficiency
of the railways.
Munuzuri et al. [13] consider that freight train terminals require complex methods of
planning management, systematically seeking to increase operational efficiency. However,
with the increase of the terminals, this management becomes more and more complex, resulting
in delays and problems of synchronism. The research proposes an algorithm-based simulation
methodology to support this management. This process has been applied in the Port Terminal
of Seville (Spain), bringing greater efficiency and manpower reduction.
Simulation models have advantages such as decision making without interfering with the
operation of the existing system, testing of new equipment without the need for large
investments, and acceleration or delay over time in phenomena that occur more slowly or fast.
They also have disadvantages, such as the need for a great experience to operate the simulation
model and the interpretation of the obtained data, the need for powerful computer resources and
extreme care with the input and output data. Due to these needs, several systems have been
developed to meet the increasingly complex modelling of rail transport, including GPSS®,
SIMAN V®, SIMSCRIPT II.5®, SLAM II®, ProModel / MedModel®, AutoMod®, Taylor II®,
WITNESS®, SIMIO® and ARENA®.
Xu et al. [14] propose a mathematical formulation to analyse a better operational efficiency
in the occupation of railway corridor lines. Xu analysed different types of models in linear
programming and concluded that the solutions may not be very efficient in a real situation and
that new studies could develop more adequate algorithms and that can reproduce more realistic
conditions.
Wang and Goverde [15] propose an energy efficiency analysis for the distribution of
schedules of a railway system through an optimization process, whose quality was verified
through several indicators (KPIs).
According to Marinov et al. [16], traffic controllers are responsible for safety, efficiency,
and trains movement on the track. The management of the system is conducted through
"Operation Plans", which define the different occurrences that the operator must execute. These
situations tend to be as broad as possible, depending on the good practices and frequencies that
they may occur, including considering the risks of a decision not being made properly.
However, the existence of a railway operation plan does not guarantee that there will be no
accidents, from rolling stock failures, line conflicts and maintenance problems. Thus, the
railway operation often requires the controllers to make decisions for conflict resolution. Such
decisions may be classified as structured, unstructured, and semi-structured, as the result of the
decision is certain and determined or uncertain and indeterminate, depending on its
characteristics. The decision also involves the following phases: Intelligence (identification of
what must be corrected), Project (identification of alternatives of conflict solutions), Choice
(identification of the best alternative) and Implementation (execution of the solution). This
demonstrates the importance of extensive analyses of the possible consequences of decision
making and that the simulation processes allow a rapid and accurate evaluation of the various
scenarios. Finally, the results of the scenarios can be expressed by indicators of technical,
economic, safety or ecological efficiency, according to the adopted criteria by the decision-
maker.
Amorim, Lopes, Silva Junior: Discrete Event-Based Railway Simulation Model for Eco-…
378
The concept of eco-efficiency was introduced by the World Bank in 1992 by [3] and has
been widely adopted worldwide. Investors have sought ways to reduce the environmental
impact of their business and at the same time increase development. The verification of eco-
efficiency is done through indicators, which must be scientifically proven, relevant
environmentally, accurate and usable for all types of business in the globe. According to the
study, the eco-efficiency index is calculated through Eq. (1), which relates the product or value
of the service to its influence on the environment.
Eco-efficiency = Product or Service Value
Environmental influence (1)
In Eq. (1), the value of the product or service generally refers to the quantity of produced
goods or services for customers, while the environmental influence refers to the consumption
of energy, water, gas emissions or any other substances harmful to the environment. Different
ventures or businesses will therefore have different ways of measuring the eco-efficiency
indicator. The principles of described eco-efficiency by the World Bank used to evaluate the
locomotives that travel on the Vitória-Minas Railroad (EFVM) [4]. An indicator used
worldwide by railroads is the Gross Ton Kilometre (GTKM) which represents the sum of the
total weight of the train multiplied by the distance. GTKM includes the weight of locomotives,
wagons and all transported freight. Since GTKM has been widely used by railroads it was used
as the service value (V) for the method.
For environmental influence, seven indicators were adopted: The total energy consumption
(E), total renewable energy consumption (RE), carbon dioxide emissions (CO2), carbon
monoxide emissions (CO), nitrogen oxides emissions (NOx), particulate matter emissions (PM)
and cost efficiency (investment, maintenance and fuel) (CE). Different scenarios representing
the exchange of fuel sources and technologies were developed, tested and analysed. The
impacts were evaluated by seven ecoefficiency performance indicators and compared with the
United States Environmental Protection Agency (EPA) standards. The results offered cost
savings and reduction opportunities.
Regarding the simulation modelling system, the present study uses AnyLogic® software,
still little known in Brazil, but widely used in the USA. This system has specific commands in
railway operation that simplify its programming in addition to allowing a greater range of
situations that can be simulated and evaluated. AnyLogic is one of the most widespread systems
on the market in terms of groups on LinkedIn as well as published papers [17].
4. DISCRETE EVENT-BASED SIMULATION MODEL FOR ECO-
EFFICIENCY EVALUATION – THE CASE OF PARANAGUÁ “KM5”
RAILYARD
4.1 The “KM5” railyard
The studied case refers to the freight railyard, also known as “KM5”, belonging to the company
Rumo Logistics, that receives wagons coming from the Paraná producing regions towards the
port terminals of the Port of Paranaguá, which is in Paranaguá City, Paraná state, Brazil. Fig. 1
shows the regional location, detail of the port and the KM5.
4.2 The input data
For the development of the KM5 simulation model, a quantitative and qualitative analysis of
the wagons, the daily schedules of the trains arriving at KM5, and the criterion of wagon
positioning were carried out.
Amorim, Lopes, Silva Junior: Discrete Event-Based Railway Simulation Model for Eco-…
379
Figure 1: Port of Paranaguá and KM5 localization [18].
The quantitative analysis considers that the operation in KM5 consists of the arrival of the
trains coming from the interior of Paraná through the mainline and that are sent to the reception
lines. Once they are stationed, railway locomotives are switched off and wagons are classified
and transferred to the parking lines using manoeuvring locomotives. After reaching several
wagons of the same customer, between 50 and 70, the locomotive draws the wagons through
the track to the cargo terminals in the port. Following, the wagons return to KM5, where they
are surveyed and then taken back to their origin.
There is great variability in the daily quantities of arriving wagons, from 160 (June 9th) to
557 (June 15th). This variability makes it difficult to operate and manage resources in the KM5.
Concerning the participation of each customer during this study period, it can be observed in
Fig. 2 the great predominance of PASA and TCP wagons and low participation of Cargill
wagons. This variation shows that while some lines are occupied and freed more often, others
take more time for this process.
Figure 2: Total share of each Costumer (June / 2016).
Amorim, Lopes, Silva Junior: Discrete Event-Based Railway Simulation Model for Eco-…
380
The qualitative wagon analysis establishes that the models of wagons used are gondola,
hopper, platform and tank types, and their specifications are indicated in Table I [19].
Table I: Wagons types.
Wagon code Wagon model Weight (t) Load
capacity (t)
Total
weight (t)
Volume
(m3) HFD Closed with hatch 19.6 60.4 80.0 75.0
HFE Hopper 23.3 76.7 100.0 114.0
HPE Hopper 22.5 77.5 100.0 100.0
TCC Tank 18.0 38.7 64.5 42.0
TCD Tank 23.5 56.5 80.0 60.0
FBD Gondola 21.2 58.8 80.0 68.0
PCC Platform 10.5 53.5 64.0 -
PCD Platform 16.0 64.0 80.0 -
PEC Platform 12.0 36.0 48.0 -
PDD Platform 16.0 64.0 80.0 -
PEC Platform 12.0 36.0 48.0 -
PED Platform 16.0 64.0 80.0 -
PND Platform 12.4 51.6 64.0 -
PPC Platform 15.5 42.0 57.5 -
PQD Platform 15.0 27.0 42.0 -
For the present study, it was assumed that all customers use HFE wagons, with a total weight
of 100 t. The wagon moving operation is done through a worksheet to control schedules and
destinations. Table II presents part of the worksheet containing information on the movement
of the wagons in the KM5, which is related to the analysis period between June and November
2016.
Table II: Composition schedule.
Train
code Product Costumer
Entrance to
KM5
KM5 to
terminal
Terminal to
KM5
Exit from
KM5 day / hour in September, 2016
K78 sugar Pasa 05 / 08:00 05 / 21:21 07 / 02:14 07 / 03:05
K78 sugar Pasa 05 / 08:00 05 / 21:21 07 / 02:14 07 / 03:05
K78 sugar Pasa 05 / 08:00 05 / 21:21 07 / 02:14 07 / 03:05
K78 sugar Pasa 05 / 08:00 05 / 21:21 07 / 02:14 07 / 03:05
K78 soybean Cotriguacu 05 / 08:00 05 / 08:40 10 / 18:21 10 / 19:20
K78 soybean Cotriguacu 05 / 08:00 05 / 08:40 10 / 04:08 10 / 05:40
K78 soybean Cotriguacu 05 / 08:00 05 / 08:40 10 / 04:08 10 / 05:40
K78 soybean Cotriguacu 05 / 08:00 05 / 08:40 10 / 18:21 10 / 19:20
As shown in Table II, each line represents the relative data of a wagon, indicating the code
of the incoming train (train code), input merchandise (product), costumer, day and time of
movements when the train arrives at the KM5 (entrance to KM5) when it leaves the KM5 and
arrives at the terminal (KM5 to the terminal) when it leaves the terminal and goes to KM5
(terminal to KM5) and when it exits KM5 back to this origin (exit from KM5). Each train
arriving at KM5 (for example K78 in Table II) is composed of a set consisting of Pasa and
Cotriguaçú wagons. The first step to use the obtained information from the rail operator Rumo
was to transform the original data format (Table II) into a format that could be interpreted by
the simulation program. In this phase an input data analysis was performed, aiming to eliminate
inconsistent data such as wagons without customer or product indication, very short time
interval between sequential trains and a very small number of wagons per train. Table III shows
the adopted solutions.
Amorim, Lopes, Silva Junior: Discrete Event-Based Railway Simulation Model for Eco-…
381
During the interview with the operators, it was reported that one part of the lines had one
defined operator, while another part was of common use. Table IV shows how lines (railways)
are organized in yard KM5 according to their occupation or use.
Table III: Problems with input data and adopted solutions.
Problem Solution
➢ Wagons without customer or product indication ➢ Removal of wagons or adoption of wagon products of the
same composition
➢ The short time interval between sequential trains ➢ The consideration that it is the same train
➢ A very small number of rail wagons per train ➢ Elimination
Table IV: Utilization of KM5´s lines.
Line Export/ Import Cargo
L1, L3, L3A, L7, L9, Uninformed Uninformed
L11, L13, L15, L17 Uninformed Train reception
L2 Export Train reception
L4, L6 Uninformed Train reception
L8 Import Grains, sugar Pasa
L10 Import Grains, sugar Bunge
L12 Import Grains, bran Cotriguaçu
L14 Import Grains, bran Cargill
L16 Import Grains, bran Dreyfus
L18 Import Cattalini + vegetable oil + empty cargo tanks
L19 Export Fertilizers
L21, L23 Export Empty wagons
L22 Import Platform wagons with a container for TCP
L24 Import Terminal circulation Line
L25 Export Containers and tank wagons
L27 Export Empty wagons
As shown in Table IV, while odd lines are used to exit the trains to the origin, the even lines
receive the destined wagons for the terminals, with certain lines having defined customers,
while others are used for other purposes.
Fig. 3 shows the schematic of KM5 and Fig. 4 shows the detail of the lines with the
nomenclature of each one of them, the even lines being for the wagons that are destined for the
port and the odd ones for the wagons coming from the port.
Figure 3: Modelling of the geometry.
From the left side of the KM5, the trains arrive from the producing regions of Paraná and
on the right, they follow towards the port terminals. In the North and the Southside are parked,
respectively, the wagons coming from the production regions and the port terminals. The model
simulates the North lines, which are named according to Table IV and highlighted by the dashed
area, as enlarged in Fig. 4.
As indicated in Table IV, there are lines for specific customers and lines without defined
destination or customers without own line (Fertipar, Fospar, Interalli, Klabin, Santa Terezinha,
Amorim, Lopes, Silva Junior: Discrete Event-Based Railway Simulation Model for Eco-…
382
Yara-KM5, Multitrans, Rocha Top-Fertilizantes, Seara, export corridor and Rumo). In this case,
a grouping of all with the fictitious name "Other" was adopted, which will be parked on a
common line.
Figure 4: Detail of the used lines by the companies.
4.3 The discrete event-based model
After inserting the geometric data of the tracks in the software AnyLogic, the rules of trains
movement and entry point in the yard are defined. to the logic is defined for the separation of
the wagons from the traction locomotive, the movement of wagon blocks, their parking and
releasing to the port terminals. The next step is to study the restrictions on the movement of
trains and wagons in terms of maximum quantity per line, passage priorities at crossing points
and paths that the locomotives should travel during manoeuvres. Finally, the scenarios of wagon
classification are simulated to identify possible restrictions of the system and proposing
solutions to eliminate them.
The AnyLogic simulation model based on discrete events considers that each event occurs
at a specific time, in this case, the arrival time of the trains to KM5. From there, it moves
locomotives and wagons following the defined operational criteria, through coupling and
decoupling movements of locomotives, as well as the defined decision-making by the system.
4.4 The classification yard strategies
After the arrival of the trains to the KM5 and liberation of the traction locomotives, the work
of the manoeuvre locomotives begins, aiming to classify the wagons in their respective lines.
This classification can be performed by two strategies. The first one (option 1), named in-block
classification, consists of the traction of all wagons blocks at the same time, that is, every
composition is drawn whole, with each group of the same customer being disconnected
sequentially in their respective line. The second strategy (option 2), called separate sorting of
wagons, consists of the manoeuvring locomotive separating groups of wagons into blocks of
the same customer, taking each block separately to its parking line, returning to fetch the next
block until all blocks are parked. Each option has advantages and disadvantages. Blocked
Amorim, Lopes, Silva Junior: Discrete Event-Based Railway Simulation Model for Eco-…
383
classification means that the manoeuvre locomotive must travel less time than the separate
classification but pulling more wagons at the same time.
4.5 Model rules
For the elaboration of the model by AnyLogic, a series of modelling rules were considered: All
wagons have 100 t of total weight, the traffic speed is equal to 5 km/h. Additionally, the model
considers that when some parking line reaches 50 wagons, they shall be pulled by the
locomotive manoeuvre to the terminal, emptying this line. Another rule refers to the maximum
number of wagons that would fit within the limits of each parking line. The purpose of this limit
is to verify the hypothesis of the total sum of the wagons already parked on the line, still, less
than 50, with wagons from a new group exceeds the physical limits of this line, parking several
wagons over the mainline. These excesses of wagons may prevent other wagons from heading
to their respective lines, causing the stoppage of the train yard. Table V shows the physical limit
of each line and the variable adopted in the modelling.
Table V: KM5 lines utilization.
Line Physical capacity
L8 78
L10 75
L12 71
L14 68
L16 65
L18 62
L22 59
L24 58
When a group of new wagons from a customer arrives at KM5, the program checks if the
number of wagons already parked on their line added to this new group exceeds the physical
capacity of the line. If this occurs, the manoeuvring locomotive receives an order to pre-empty
the line, freeing up space for the new wagons to be parked.
The third rule solves a specific situation of the blocked classification of the wagons into the
parking lot. If, for example, a train arrives at KM5 with a group of wagons such that the first
group of wagons to be parked contains more than the capacity of its line, this first group may
leave wagons invading the line which serves as another parking area, causing a conflict. The
solution was to adopt, in the programming logic, prior verification of this situation and, when
this happens, the model temporarily changes the positioning option of the wagons, ungrouping
them and performing the separate classification. The next train that arrives will resume the
blocked classification.
Figure 5: Crossing point on the tracks.
Amorim, Lopes, Silva Junior: Discrete Event-Based Railway Simulation Model for Eco-…
384
Regarding the safety of crossings, the program tried to perform a signalling analysis at the
points of conflict. Fig. 5 shows the crossing of lines between the manoeuvring locomotive and
the traction locomotives. The simulation model predicts that the crossover point can only be
accessed if it is free. It can be observed that the two traction locomotives can pass at the
intersection. The locomotive must wait for its clearance.
4.6 Eco-efficiency evaluation
To evaluate the eco-efficiency of the proposed classification strategies, two simulations were
carried out for the month of June 2016. The maneuverer locomotive movement data was
obtained, such as the total displacement time of the maneuverer locomotive in hours, the
distance travelled (km) and GTKM (ton × km) as shown in Table VI. These values served as
parameters for the calculation of eco-efficiency indexes.
Table VI: Scenarios summary.
Scenario Unity Option 1 Option 2
Time h 219.84 252.71
Distance km 1,099.18 1,263.55
GTKM ton × km 3,743,310.56 3,068,969.51
The calculated efficiency ratios are total energy consumption (E), Carbon dioxide emissions
(CO2), Carbon monoxide emissions (CO), Nitrogen oxides emissions (NOX), Particulate matter
emissions (PM) and Total cost (investment, maintenance and fuel) (CE).
The following technical parameters were used [18]: Fuel consumption (FE) = 292.91 l / h;
Calorie content of fuel (CC) = 37.711 kJ / l; Cost of fuel (FU) = US $ 1.09 / l; CO2 fuel emission
factor (EC) = 2.71 kg / l; CO fuel emission factor (EO) = 4.5 g / l; NOX fuel emission factor
(EN) = 44.3 g / l; Particulate matter fuel emission factor (EP) = 1.62 g / l.
The total renewable energy consumption (eiRE), the MC (the maintenance cost) and IC
(investment cost) were not included in this analysis. Table VII shows the results of simulated
classification strategies.
Table VII: Summary of classification strategies.
Scenario Unit Option 1 Option 2
Calculated efficiency
eiE kJ 2,428,296.53 2,791,416.72
eiCO2 Kg 173,859.10 199,857.47
eiCO kg 37.31 42.88
eiNOX kg 2,852.58 3,279.14
eiPM kg 104.32 119.91
eiCE US$ 70,187.56 80,683.20
Eco-efficiency index
E ton × km / kJ 1.54 1.10
RE ton × km / kg 21.53 15.36
CO2 ton × km / kg 100,340.40 71,563.17
NOX ton × km / kg 1,312.26 935.91
PM ton × km / kg 35,884.52 25,592.98
CE ton × km / US$ 53.33 38.04
Based on the obtained results, it can be observed that option 1 is more advantageous than
option 2. Option 1 required 13 % less time and generates an accumulated transportation moment
22 % higher. This translates into less cost of the use of the locomotives and higher freight
handling. Regarding eco-efficiency indexes, option 1 has 40 % higher indexes than option 2.
This shows that less environmental impact is caused by moving all the wagon blocks at the
same time than by cutting the composition into separate blocks. The simulations carried out in
this paper were made available for online execution through the electronic addresses.
Amorim, Lopes, Silva Junior: Discrete Event-Based Railway Simulation Model for Eco-…
385
Option 1 simulation: https://bit.ly/3gbJ6tF
Option 2 simulation: https://bit.ly/3gdoSQl
5. CONCLUSIONS
This paper demonstrates that several aspects need to be considered to evaluate the
environmental impact of on railway operation. Not only the fuel consumption itself but how
efficiently it is being consumed. This new methodology of analysis can help to see how different
strategies of operation can influence the results of the railyard management.
The interviews with rail operators demonstrated that different situations require different
decisions and often uncommon ones at that. In this context, extensive knowledge of the rules
of wagon movement is fundamental so that the model can represent the operation in a train yard
more accurately. The installation of railyard surveillance cameras would enable the observation
of daily events and movement, also helping in the informed definition of programming logic.
AnyLogic is very robust and efficient software for modelling and simulation. Its specific
functions for railway evaluation facilitate the development of ever more complex models,
allowing a wide range of situations to be reproduced.
The analysis of the eco-efficiency indexes can be expanded considering different models of
wagons, with different capacities, dimensions and weights, different classes of locomotives and
the types of cargo that are transported. Different solutions for engines and their fuel can also be
evaluated using this process.
The application of a simulation model for the analysis of eco-efficiency is quite efficient
since it allows the rapid surveying of cargo movement. Moreover, this method allows other
studies such as the exit operation of the wagons back to their origin and their operation in the
region of the port terminals. This knowledge allows for the development of manuals and wider
operation standards, including situations and problems not precisely evaluated with such
precision by traditional methods. Besides, different models of locomotives and wagons can be
compared to identify the ones that bring more efficiency to the system.
ACKNOWLEDGEMENTS
The present study was carried out with the support of the Coordination of Improvement of Higher
Education Personnel - Brazil (CAPES, Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
- Brasil) - Financing Code 001. To the professors Renata Albergaria and Paulo Afonso Lopes for the
orientations, revisions and collaboration for this paper.
REFERENCES
[1] AnyLogic. Simulation software, from http://www.anylogic.com/, accessed on 03-05-2018
[2] Chwif, L.; Medina, A. C. (2006). Modelagem e Simulações de Eventos Discretos – Teoria &
Aplicações, Palas Athena, São Paulo
[3] Verfaillie, H. A.; Bidwell, R. (2000). Measuring Eco-Efficiency. A Guide to Reporting Company
Performance, World Business Council for Sustainable Development, Geneva
[4] Carvalhaes, B. B.; Rosa, R. D. A.; D´Agosto, M. D. A.; Ribeiro, G. M. (2017). A method to
measure the eco-efficiency of diesel locomotive, Transportation Research Part D: Transport and
Environment, Vol. 51, 29-42, doi:10.1016/j.trd.2016.11.031
[5] Fan, W. G.; Hou, G. Y.; Wang, W. X.; Zhang, X. L.; Wang, J. D. (2019). Design and dynamic
analysis of a new rail grinding device using closed abrasive belt, International Journal of
Simulation Modelling, Vol. 18, No. 3, 531-542, doi:10.2507/IJSIMM18(3)CO14
[6] Wang, Y. R.; Chen, A. N. (2016). Production logistics simulation and optimization of industrial
enterprise based on Flexsim, International Journal of Simulation Modelling, Vol. 15, No. 4, 732-
741, doi:10.2507/IJSIMM15(4)CO18
Amorim, Lopes, Silva Junior: Discrete Event-Based Railway Simulation Model for Eco-…
386
[7] Chen, Y. X. (2016). Integrated optimization model for production planning and scheduling with
logistics constraints, International Journal of Simulation Modelling, Vol. 15, No. 4, 711-720,
doi:10.2507/IJSIMM15(4)CO16
[8] Suryani, E.; Hendrawan, R. A.; Adipraja, P. F. E.; Indraswari, R. (2020). System dynamics
simulation model for urban transportation planning: a case study, International Journal of
Simulation Modelling, Vol. 19, No. 1, 5-16, doi:10.2507/IJSIMM19-1-493
[9] Fragapane, G. I.; Zhang, C.; Sgarbossa, F.; Strandhagen, J. O. (2019). An agent-based simulation
approach to model hospital logistics, International Journal of Simulation Modelling, Vol. 18, No.
4, 654-665, doi:10.2507/IJSIMM18(4)497
[10] Confessore, G.; Liotta, G.; Cicini, P.; Rondinome, F.; De Luca, P. (2009). A simulation-based
approach for estimating the commercial capacity of railways, Proceedings of the 2009 Winter
Simulation Conference, 2542-2552, doi:10.1109/WSC.2009.5429664
[11] Crane, R. R.; Brown, F. B.; Blanchard, R. O. (1955). An analysis of a railroad classification yard,
Journal of the Operation Research Society of America, Vol. 3, No. 3, 262-271,
doi:10.1287/opre.3.3.262
[12] Assad, A. A. (1980). Models for rail transportation, Transportation Research Part A: General,
Vol. 14, No. 3, 205-220, doi:10.1016/0191-2607(80)90017-5
[13] Munuzuri, J.; Domingueza, I.; Berrocal, M. A.; Escudero, A. (2016). An allocation-scheduling
heuristic to manage train traffic in an intermodal terminal, Computers in Industry, Vol. 82, 196-
204, doi:10.1016/j.compind.2016.07.006
[14] Xu, Y.; Jia, B.; Ghiasi, A.; Li, X. (2017). Train routing and timetabling problem for heterogeneous
train traffic with switchable scheduling rules, Transportation Research Part C: Emerging
Technologies, Vol. 84, 196-218, doi:10.1016/j.trc.2017.08.010
[15] Wang, P.; Goverde, R. M. P. (2019). Multi-train trajectory optimization for energy-efficient
timetabling, European Journal of Operational Research, Vol. 272, No. 2, 621-635,
doi:10.1016/j.ejor.2018.06.034
[16] Marinov, M.; Sahin, I.; Ricci, S.; Vasic-Franklin, G. (2013). Railway operations, time-tabling and
control, Research in Transportation Economics, Vol. 41, No. 1, 59-75, doi:10.1016/
j.retrec.2012.10.003
[17] GENOA. Anylogic para Simulação de Porto, Rodovias e Ferrovias, from
http://www.genoads.com.br, accessed on 19-03-2019
[18] Prefeitura de Paranaguá. Paranagua Port, from http://www.paranagua.pr.gov.br, accessed on 19-
03-2019
[19] Centro-Oeste. Brazilian Rail History, from http://vfco.brazilia.jor.br, accessed on 28-02-2019